【作者】 邱勇；

【导师】 张维；

【作者基本信息】 天津大学 ， 管理科学与工程， 2007， 博士

【摘要】 通过构建最优的多银行中小企业贷款池,运用合成CDO等结构化融资方式为化解中小企业信贷资产风险、缓解中小企业融资困境提供了新思路。这类创新机制在一定程度上避免了由信息不对称导致贷款资产再出售过程中的逆向选择行为,减少了传统证券化方式的效率损失。该创新机制将有效提高中小企业贷款资产的流动性,降低商业银行信贷资产风险集中度和监管资本成本,有利于扩大商业银行对中小企业的贷款供给,有利于缓解中小企业“贷款难、融资难”的困境。本文首先设计提出了基于多银行贷款池的中小企业贷款合成CDO交易结构。该交易结构是多种信用衍生产品和风险管理技术的复合,在减少传统证券化模式的效率损失的基础上,为化解中小企业信贷资产的潜在风险、缓解中小企业融资困境提出了创新机制。多银行的中小企业信贷资产组合的合理分档是顺利实施以上合成CDO交易结构的重要前提条件。本文运用证券拍卖博弈理论,首先在两类投资者的市场假设条件下,得到了投资者参与分档证券发售的混合策略均衡,以及第二类投资者参与竞拍的条件,根据投资者老练程度和资产组合质量,发行机构通过两层分档吸引两类投资者,能够最大化其期望收益;在多类投资者的一般情形下,首先得到了所有类型的投资者都参与该证券拍卖的均衡解,及其参与证券发售的条件,发行人通过多层分档,吸引尽量多的投资者参与竞拍,能够最大化其预期收益。参考资产组合的违约风险特征和相关性结构是基于多银行贷款池的合成CDO分档证券定价的关键问题。本文在考虑各银行风险因素的基础上,构建了反映多银行贷款组合风险特征和相关性结构的多因素Copula模型,并在条件独立性假设下得到了该贷款资产组合的无条件违约概率分布规律;在多元正态分布Copula模型和多元正态逆高斯分布Copula模型下,分别研究了该贷款组合的违约风险特征和相关性结构,正态逆高斯分布不仅能够比较清楚刻画本文所研究的多银行贷款组合的违约风险特征,而且能够很好地刻画组合违约风险的厚尾分布特性。在无套利均衡的合成CDO的公允价差定价框架下,参考贷款资产组合在各付息时点的期望违约损失是合理定价的另一关键,本文分别在同质的贷款资产组合和异质的贷款资产组合情形下,得到了多银行贷款资产组合期望违约损失以及合成CDO分档的合理定价,反映了该合成CDO产品的风险特征和优先结构。更多还原

【Abstract】 The purpose of this dissertation is to design an innovative mechanism for portfolio risk management of SME loans. Synthetic CDOs based on Multi-bank Loan Pool are designed and employed to diversify the credit risk of SME loans, and relieve the dilemma of SME financing. The mechanism can to some extent avoid the adverse selection caused by the information asymmetry in the process of SME loan resale, and reduce the efficiency loss of traditional securitization. Its implementation can effectively improve the liquidity of SME loans, and decrease the risk convergence and the costs of regulatory capital requirements for commercial banks. As a result, it will increase the supply of SME loans, and relieve the dilemma of SME financing.Firstly the transaction structure of SME loan synthetic CDOs based on Multi-bank Loan Pool is designed. The structure is composed of some credit derivatives and risk management methods. The innovative mechanism can effectively reduce the efficiency loss of traditional securitization, and relieve the dilemma of SME financing.Tranching of Multi-bank Loan Pool is the important premise to carry out the above structured mechanism. Auction Game is applied to derive the equilibrium of tranching security private issuance. Under the condition of two kinds of investors, a unique mixed strategy equilibrium is obtained for investors’auction of tranching security. According to the investors’sophistication and the quality of asset portfolio, the auctioning conditions of the second type of investors are derived. Issuers can maximize their expected revenue by issuing two tranches. Under the condition of more than two kinds of investors, a unique equilibrium is also obtained for investors’auction; issuers absorb numerous investors to action the tranches and maximize their expected revenue with more tranches.The characteristics and dependence structure of reference asset portfolio is one of keys for synthetic CDO tranche pricing. Multi-bank risk factors are considered in this thesis; a multi-factor Copula model is constructed to reflect the portfolio risk and dependence structure of Multi-bank Loan Pool designed above. Unconditional default probability distribution is derived under the conditional independence, the multivariate Normal Gaussian Copula model and multivariate Normal Inverse Gaussian Copula model are employed to investigate the characteristics of reference portfolio’s default risk. The results show that the multivariate Normal Inverse Gaussian Copula model can not only reflect the default risk characteristics, but also catch the tail-fatness of reference loan portfolio.Under the no-arbitrage pricing framework of synthetic CDOs, expected default loss at each spread payment dates is the another key for tranche pricing. Homogeneous and heterogeneous loan portfolios are considered in this thesis. Expected default loss of Multi-bank Loan Pool and synthetic CDO tranche fair spreads are derived respectively. The results show that the above tranche pricing model can reflect the risk characteristics and seniority structure of synthetic CDOs designed in this thesis based on Multi-bank SME loan pool.更多还原